JFET that ain't no JFET...

[Rob Strand]

Since there's some JFETs that performs better than others, is there any method for choosing "good ones"?

Cheers!

To be true to the standard RAT circuit the JFET source should be biased at 1V to 2V.  Perhaps some preference keeping it over 1.5V.

However ...   many people have stated it sounds better biased somewhat higher bias voltage.   A good example of a circuit which does this forcefully is the ggg rodent circuit.    That biases the gate at 2.8V and I think that ends up with the JFET source at around 4V or 5V.
http://www.generalguitargadgets.com/pdf/ggg_rodent_instruct.pdf
http://www.generalguitargadgets.com/pdf/ggg_rodent_sc.pdf

One of the later RATs biases the JFETs with the gates connected to 4.5V via a resistor, higher again than the ggg rodent.  BTW, that's pretty much the circuit you have for the BJT except with a JFET in there!   That's going to bias the JFET source at say 5V, perhaps even higher.   Quite different biasing to the standard RAT.

So yes a lot of variations!

Your BJT is probably sitting at a 2.5V to 3V.   So not bad.    You could do a lot worse with a bad JFET.   The problem with the standard RAT circuit is it will bias the JFET poorly if the JFET's VP (or Vgs_off) parameter is low.   Unfortunately a lot of JFETs sold now have low VP (or Vgs_off) parameters.

[fowl]

There are a few odd cap values

There is no oddness in E6 series values..

Well, I had never kept any 150n caps around, not hard to find but not very commonly used either.  I ended up paralleling two little ceramics to fit in the space afforded by the layout I was following.  Now I have ordered some, for the next build I guess.

To be true to the standard RAT circuit the JFET source should be biased at 1V to 2V.  Perhaps some preference keeping it over 1.5V.

This has been my experience.  The distortion gets too muddy to me below maybe 1.2V or so.  I select for 1.5 or higher, I don't hear any difference above that.

^edit^:  Just thought later about the relationship between that source voltage in the Rat, and the forward voltage of the clipping diodes right before it in the circuit.  I've always built them with regular silicon diodes, one each direction.  Maybe with LED or asymmetrical clippers you would want to look for something more like 2.5V for the headroom.

[Rob Strand]

Using MOSFET buffers as alternatives to JFET buffers isn't so popular in the community.    The behaviour should be closer to JFETs than a BJT, especially the high input impedance.   The downside is the common parts tend to be noisier.   

[Steben]

As said: few perfect jFETs are around for that specific use.
If you need to bias them like a BJT, it is better IMHO to use a BJT. Nothing as solid as a BJT emitter follower. It is one of those building blocks I never ever had issues with.

Rob, you mention low Vgs offerings. These are simply better for gainy applications rather than buffers, not?

[antonis]

>off-topic ON<

Vgs is another sob story..
(especially for power supply units - in the mean of, for a power BJT you "loose" about 1 Volt where for a power Mosfet you "loose" about 5V..)
Of course, in case of no problem with total supply voltage, verdict is in favour of Mosfets for many assets..)

>off-topic OFF<

P.S.
Plz be positive about a guy with 12 blown out IRF3205 only a short while ago..

[anotherjim]

I reverse faked a JFET tonight. What I thought was a 2N3904 in the board I was populating was a genuine Fairchild 2N5457. When you build from a breadboard - check you pulled the right part to solder in!

[antonis]

When you build from a breadboard - check you pulled the right part to solder in!

Never ever solder anything previously used on breadboard..!!

[fowl]

Never ever solder anything previously used on breadboard..!!

wat?

[anotherjim]

I clean all the flour and butter off my breadboards first!